Simulation and Forecasting of Floods based on Asymmetric Laplace Unit Hydrograph Model ( case study : sub-basin of Jahrom, Fars province, Iran)

In recent years, the occurrence of malicious floods in our country has caused many financial and financial losses. In February 2017, a huge flood occurred in Jahrom city (Fars province, Iran) which broke the dam and caused many problems for the people and the city administrators. Simulation and prediction of severity and time of occurrence of floods can prevent these losses largely and it can provide suitable information for use in flood alert systems. For this purpose, rainfall analysis and simulation of observation flood performed by optimizing and calibrating the parameters of the rainfall-runoff model as the Asymmetric Laplace Unit Hydrograph(ALUH) based on the Nash-Sutcliffe index. Calibration, verification and evaluation of performance of models according to Nash-Sutcliffe efficiency (NSE) and Root mean square error of observations to Standard deviation Ratio (RSR) indices performed in 10 selected events. Then, empirical relationships between characteristics of rainfall and runoff and sensitivity of ALUH model to each parameter investigated. For modeling, precipitation data and satellite data from Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks (PERSIANN CCS) used as inputs for two hydrometry stations (Babaarab and Sarvo) located in the sub-basin of Jahrom watershed. The results indicate that regional precipitation is mainly in form of rain and more than 90 mm/day rainfall and 700 m³/s peak flow and a total annual rainfall of 750 (mm/year) is possible with return period of 100 years .

The simulation results show that the model in the calibration stage for both of the two modes of using precipitation data and PERSIANN CCS, respectively, with an average Nash-Sutcliff index of 0.8 and 0.71 for the Babaarab station, and the mean of the Nash Sutcliff index 0.82 and 0.67 have been very good for the Sarvo station. The results of the sensitivity analysis on the model parameters indicate that the decrease in the change in the parameter of the peak-time event can be 20 times more effective than the change in the other two parameters in the Nash Sutcliff model.

Keywords: Rainfall-runoff model, Asymmetric Laplace Unit Hydrograph, Calibration, Validation, Nash-Sutcliffe index.